nfsd performs abysmally on this machine under conditions in which Solaris's NFS implementation is reasonably fast, and while local IO to the same filesystems is still zippy. For instance, copying a 4GB file over NFSv3 from a ZFS filesystem with the following flags [rw,nosuid,hard,intr,nofsc,tcp,vers=3,rsize=8192,wsize=8192,sloppy,addr=X.X.X.X](Linux client, the above is the server), I achieve 2 MB/s, fluctuating between 1 and 3. (pv reports 2.23 MB/s avg) Locally, on the server, I achieve 110-140 MB/s (at the end of pv, it reports 123 MB/s avg). I'd assume network latency, but nc with no flags other than port achieves 30-50 MB/s between server and client. Latency is also abysmal - ls on a randomly chosen homedir full of files, according to time, takes: real 0m15.634s user 0m0.012s sys 0m0.097s while on the local machine: real 0m0.266s user 0m0.007s sys 0m0.000s The server in question is a 3GHz Core 2 Duo, running FreeBSD RELENG_8. The kernel conf, DTRACE_POLL, is just the stock AMD64 kernel with all of the DTRACE-related options turned on, as well as the option to enable polling in the NIC drivers, since we were wondering if that would improve our performance. We tested this with a UFS directory as well, because we were curious if this was an NFS/ZFS interaction - we still got 1-2 MB/s read speed and horrible latency while achieving fast throughput and latency local to the server, so we're reasonably certain it's not "just" ZFS, if there is indeed any interaction there. Read speed of a randomly generated 6500 MB file on UFS over NFSv3 with the same flags as above: 1-3 MB/s, averaging 2.11 MB/s Read speed of the same file, local to the server: consistently between 40-60 MB/s, averaging 61.8 MB/s [it got faster over time - presumably UFS was aggressively caching the file, or something?] Read speed of the same file over NFS again, after the local test: Amusingly, worse (768 KB/s-2.2 MB/s, with random stalls - average reported 270 KB/s(!)). How-To-Repeat: 1) Mount multiple NFS filesystems from the server 2) Watch as your operations latency and throughput rapidly sink to near-zero
On Tue, 30 Mar 2010, Rich Ercolani wrote: >> Description: > nfsd performs abysmally on this machine under conditions in which Solaris's NFS implementation is reasonably fast, and while local IO to the same filesystems is still zippy. Please don't format lines for 200+ column terminals. Does it work better when limited to 1 thread (nfsd -n 1)? In at least some versions of it (or maybe in nfsiod), multiple threads fight each other under load. > For instance, copying a 4GB file over NFSv3 from a ZFS filesystem with the following flags [rw,nosuid,hard,intr,nofsc,tcp,vers=3,rsize=8192,wsize=8192,sloppy,addr=X.X.X.X](Linux client, the above is the server), I achieve 2 MB/s, fluctuating between 1 and 3. (pv reports 2.23 MB/s avg) > > Locally, on the server, I achieve 110-140 MB/s (at the end of pv, it reports 123 MB/s avg). > > I'd assume network latency, but nc with no flags other than port achieves 30-50 MB/s between server and client. > > Latency is also abysmal - ls on a randomly chosen homedir full of files, according to time, takes: > real 0m15.634s > user 0m0.012s > sys 0m0.097s > while on the local machine: > real 0m0.266s > user 0m0.007s > sys 0m0.000s It probably is latency. nfs is very latency-sensitive when there are lots of small files. Transfers of large files shouldn't be affected so much. > The server in question is a 3GHz Core 2 Duo, running FreeBSD RELENG_8. The kernel conf, DTRACE_POLL, is just the stock AMD64 kernel with all of the DTRACE-related options turned on, as well as the option to enable polling in the NIC drivers, since we were wondering if that would improve our performance. Enabling polling is a good way to destroy latency. A ping latency of more that about 50uS causes noticable loss of performance for nfs, but LAN latency is usually a few times higher than that, and polling without increasing the clock interrupt frequency to an excessively high value gives a latency of at least 20 times higher than that. Also, -current with debugging options is so bloated that even localhost has a ping latency of about 50uS on a Core2 (up from 2uS for FreeBSD-4 on an AthlonXP). Anyway try nfs on localhost to see if reducing the latency helps. > We tested this with a UFS directory as well, because we were curious if this was an NFS/ZFS interaction - we still got 1-2 MB/s read speed and horrible latency while achieving fast throughput and latency local to the server, so we're reasonably certain it's not "just" ZFS, if there is indeed any interaction there. After various tuning and bug fixing (now partly committed by others) I get improvements like the following on low-end systems with ffs (I don't use zfs): - very low end with 100Mbps ethernet: little change; bulk transfers always went at near wire speed (about 10 MB/S) - low end with 1Gbps/S: bulk transfers up from 20MB/S to 45MB/S (local ffs 50MB/S). buildworld over nfs of 5.2 world down from 1200 seconds to 800 seconds (this one is very latency-sensitive. Takes about 750 seconds on local ffs). > Read speed of a randomly generated 6500 MB file on UFS over NFSv3 with the same flags as above: 1-3 MB/s, averaging 2.11 MB/s > Read speed of the same file, local to the server: consistently between 40-60 MB/s, averaging 61.8 MB/s [it got faster over time - presumably UFS was aggressively caching the file, or something?] You should use a file size larger than the size of main memory to prevent caching, especially for reads. That is 1GB on my low-end systems. > Read speed of the same file over NFS again, after the local test: > Amusingly, worse (768 KB/s-2.2 MB/s, with random stalls - average reported 270 KB/s(!)). The random stalls are typical of the problem with the nfsd's getting in each other's way, and/or of related problems. The stalls that I saw were very easy to see in real time using "netstat -I <interface> 1" -- they happened every few seconds and lasted a second or 2. But they were never long enough to reduce the throughput by more than a factor of 3, so I always got over 19 MB/S. The throughput was reduced by approximately the ratio of stalled time to non-stalled time. >> How-To-Repeat: > 1) Mount multiple NFS filesystems from the server > 2) Watch as your operations latency and throughput rapidly sink to near-zero Multiple active nfs mounts are probably a different problem. You certainly need more than 1 nfsd and/or nfsiod to handle them, and the stalls might be a result of not having enough daemons. Bruce
On Tue, Mar 30, 2010 at 11:50 AM, Bruce Evans <brde@optusnet.com.au> wrote: > Does it work better when limited to 1 thread (nfsd -n 1)? =A0In at least > some versions of it (or maybe in nfsiod), multiple threads fight each oth= er > under load. It doesn't seem to - nfsd -n 1 still ranges between 1-3 MB/s for files > RAM on server or client (6 and 4 GB, respectively). >> For instance, copying a 4GB file over NFSv3 from a ZFS filesystem with t= he >> following flags >> [rw,nosuid,hard,intr,nofsc,tcp,vers=3D3,rsize=3D8192,wsize=3D8192,sloppy= ,addr=3DX.X.X.X](Linux >> client, the above is the server), I achieve 2 MB/s, fluctuating between = 1 >> and 3. (pv reports 2.23 MB/s avg) >> >> Locally, on the server, I achieve 110-140 MB/s (at the end of pv, it >> reports 123 MB/s avg). >> >> I'd assume network latency, but nc with no flags other than port achieve= s >> 30-50 MB/s between server and client. >> >> Latency is also abysmal - ls on a randomly chosen homedir full of files, >> according to time, takes: >> real =A0 =A00m15.634s >> user =A0 =A00m0.012s >> sys =A0 =A0 0m0.097s >> while on the local machine: >> real =A0 =A00m0.266s >> user =A0 =A00m0.007s >> sys =A0 =A0 0m0.000s > > It probably is latency. =A0nfs is very latency-sensitive when there are l= ots > of small files. =A0Transfers of large files shouldn't be affected so much= . Sure, and next on my TODO is to look into whether 9.0-CURRENT makes certain ZFS high-latency things perform better. >> The server in question is a 3GHz Core 2 Duo, running FreeBSD RELENG_8. T= he >> kernel conf, DTRACE_POLL, is just the stock AMD64 kernel with all of the >> DTRACE-related options turned on, as well as the option to enable pollin= g in >> the NIC drivers, since we were wondering if that would improve our >> performance. > > Enabling polling is a good way to destroy latency. =A0A ping latency of > more that about 50uS causes noticable loss of performance for nfs, but > LAN latency is usually a few times higher than that, and polling without > increasing the clock interrupt frequency to an excessively high value > gives a latency of at least 20 times higher than that. =A0Also, -current > with debugging options is so bloated that even localhost has a ping > latency of about 50uS on a Core2 (up from 2uS for FreeBSD-4 on an > AthlonXP). =A0Anyway try nfs on localhost to see if reducing the latency > helps. Actually, we noticed that throughput appeared to get marginally better whil= e causing occasional bursts of crushing latency, but yes, we have it on in th= e kernel without using it in any actual NICs at present. :) But yes, I'm getting 40-90+ MB/s, occasionally slowing to 20-30 MB/s, average after copying a 6.5 GB file of 52.7 MB/s, on localhost IPv4, with no additional mount flags. {r,w}size=3D8192 on localhost goes up to 80-100 MB/s, with occasional sinks to 60 (average after copying another, separate 6.5 GB file: 77.3 MB/s). Also: 64 bytes from 127.0.0.1: icmp_seq=3D0 ttl=3D64 time=3D0.015 ms 64 bytes from 127.0.0.1: icmp_seq=3D1 ttl=3D64 time=3D0.049 ms 64 bytes from 127.0.0.1: icmp_seq=3D2 ttl=3D64 time=3D0.012 ms 64 bytes from [actual IP]: icmp_seq=3D0 ttl=3D64 time=3D0.019 ms 64 bytes from [actual IP]: icmp_seq=3D1 ttl=3D64 time=3D0.015 ms >> We tested this with a UFS directory as well, because we were curious if >> this was an NFS/ZFS interaction - we still got 1-2 MB/s read speed and >> horrible latency while achieving fast throughput and latency local to th= e >> server, so we're reasonably certain it's not "just" ZFS, if there is ind= eed >> any interaction there. > > After various tuning and bug fixing (now partly committed by others) I ge= t > improvements like the following on low-end systems with ffs (I don't use > zfs): > - very low end with 100Mbps ethernet: little change; bulk transfers alway= s > =A0went at near wire speed (about 10 MB/S) > - low end with 1Gbps/S: bulk transfers up from 20MB/S to 45MB/S (local ff= s > =A050MB/S). =A0buildworld over nfs of 5.2 world down from 1200 seconds to= 800 > =A0seconds (this one is very latency-sensitive. =A0Takes about 750 second= s on > =A0local ffs). Is this on 9.0-CURRENT, or RELENG_8, or something else? >> Read speed of a randomly generated 6500 MB file on UFS over NFSv3 with t= he >> same flags as above: 1-3 MB/s, averaging 2.11 MB/s >> Read speed of the same file, local to the server: consistently between >> 40-60 MB/s, averaging 61.8 MB/s [it got faster over time - presumably UF= S >> was aggressively caching the file, or something?] > > You should use a file size larger than the size of main memory to prevent > caching, especially for reads. =A0That is 1GB on my low-end systems. I didn't mention the server's RAM, explicitly, but it has 6 GB of real RAM, and the files used were 6.5-7 GB each in that case (I did use a 4GB file earlier - I've avoided doing that again here). >> Read speed of the same file over NFS again, after the local test: >> Amusingly, worse (768 KB/s-2.2 MB/s, with random stalls - average report= ed >> 270 KB/s(!)). > > The random stalls are typical of the problem with the nfsd's getting > in each other's way, and/or of related problems. =A0The stalls that I > saw were very easy to see in real time using "netstat -I <interface> > 1" -- they happened every few seconds and lasted a second or 2. =A0But > they were never long enough to reduce the throughput by more than a > factor of 3, so I always got over 19 MB/S. =A0The throughput was reduced > by approximately the ratio of stalled time to non-stalled time. I believe it. I'm seeing at least partially similar behavior here, when I mention the performance drops where transfer briefly pauses and then picks up again in the localhost case, even with nfsd -n 1 and nfsiod -n 1. - Rich
On Tue, 30 Mar 2010, Rich wrote: > On Tue, Mar 30, 2010 at 11:50 AM, Bruce Evans <brde@optusnet.com.au> wrote: >>> For instance, copying a 4GB file over NFSv3 from a ZFS filesystem with the >>> following flags >>> [rw,nosuid,hard,intr,nofsc,tcp,vers=3,rsize=8192,wsize=8192,sloppy,addr=X.X.X.X](Linux >>> client, the above is the server), I achieve 2 MB/s, fluctuating between 1 >>> and 3. (pv reports 2.23 MB/s avg) I also tried various nfs r/w sizes and tcp/udp. The best sizes are probably the fs block size or twice that (normally 16K for ffs). Old versions of FreeBSD had even more bugs in this area and gave surprising performance differences depending on the nfs r/w sizes or application i/o sizes. In some cases smaller sizes worked best, apparently because they avoided the stalls. >>> ... >> Enabling polling is a good way to destroy latency. A ping latency of >> ... > Actually, we noticed that throughput appeared to get marginally better while > causing occasional bursts of crushing latency, but yes, we have it on in the > kernel without using it in any actual NICs at present. :) > > But yes, I'm getting 40-90+ MB/s, occasionally slowing to 20-30 MB/s, > average after copying a 6.5 GB file of 52.7 MB/s, on localhost IPv4, > with no additional mount flags. {r,w}size=8192 on localhost goes up to > 80-100 MB/s, with occasional sinks to 60 (average after copying > another, separate 6.5 GB file: 77.3 MB/s). I thought you said you often got 1-3MB/S. > Also: > 64 bytes from 127.0.0.1: icmp_seq=0 ttl=64 time=0.015 ms > 64 bytes from 127.0.0.1: icmp_seq=1 ttl=64 time=0.049 ms > 64 bytes from 127.0.0.1: icmp_seq=2 ttl=64 time=0.012 ms Fairly normal slowness for -current. > 64 bytes from [actual IP]: icmp_seq=0 ttl=64 time=0.019 ms > 64 bytes from [actual IP]: icmp_seq=1 ttl=64 time=0.015 ms Are these with external hardware NICs? Then 15 uS is excellent. Better than I've ever seen. Very good hardware might be able to do this, but I suspect it is for the local machine. BTW, I don't like the times been reported in ms and sub-uS times not being supported. I sometimes run Linux or cygwin ping and don't like it not supporting sub-mS times, so that it always reports 0 for my average latency of 100 uS. >> After various tuning and bug fixing (now partly committed by others) I get >> improvements like the following on low-end systems with ffs (I don't use >> zfs): >> - very low end with 100Mbps ethernet: little change; bulk transfers always >> went at near wire speed (about 10 MB/S) >> - low end with 1Gbps/S: bulk transfers up from 20MB/S to 45MB/S (local ffs >> 50MB/S). buildworld over nfs of 5.2 world down from 1200 seconds to 800 >> seconds (this one is very latency-sensitive. Takes about 750 seconds on >> local ffs). > > Is this on 9.0-CURRENT, or RELENG_8, or something else? Mostly with 7-CURRENT or 8-CURRENT a couple of years ago. Sometimes with a ~5.2-SERVER. nfs didn't vary much with the server, except there were surprising differences due to latency that I never tracked down. I forgot to mention another thing you can try easily: - negative name caching. Improves latency. I used this to reduce makeworld times significantly, and it is now standard in -current but not enabled by default. Bruce
On Tue, Mar 30, 2010 at 1:11 PM, Bruce Evans <brde@optusnet.com.au> wrote: > On Tue, 30 Mar 2010, Rich wrote: > >> On Tue, Mar 30, 2010 at 11:50 AM, Bruce Evans <brde@optusnet.com.au> >> wrote: > >>>> For instance, copying a 4GB file over NFSv3 from a ZFS filesystem with >>>> the >>>> following flags >>>> >>>> [rw,nosuid,hard,intr,nofsc,tcp,vers=3D3,rsize=3D8192,wsize=3D8192,slop= py,addr=3DX.X.X.X](Linux >>>> client, the above is the server), I achieve 2 MB/s, fluctuating betwee= n >>>> 1 >>>> and 3. (pv reports 2.23 MB/s avg) > > I also tried various nfs r/w sizes and tcp/udp. =A0The best sizes are > probably the fs block size or twice that (normally 16K for ffs). =A0Old > versions of FreeBSD had even more bugs in this area and gave surprising > performance differences depending on the nfs r/w sizes or application > i/o sizes. =A0In some cases smaller sizes worked best, apparently because > they avoided the stalls. > >>>> ... >>> >>> Enabling polling is a good way to destroy latency. =A0A ping latency of >>> ... > >> Actually, we noticed that throughput appeared to get marginally better >> while >> causing occasional bursts of crushing latency, but yes, we have it on in >> the >> kernel without using it in any actual NICs at present. :) >> >> But yes, I'm getting 40-90+ MB/s, occasionally slowing to 20-30 MB/s, >> average after copying a 6.5 GB file of 52.7 MB/s, on localhost IPv4, >> with no additional mount flags. {r,w}size=3D8192 on localhost goes up to >> 80-100 MB/s, with occasional sinks to 60 (average after copying >> another, separate 6.5 GB file: 77.3 MB/s). > > I thought you said you often got 1-3MB/S. > >> Also: >> 64 bytes from 127.0.0.1: icmp_seq=3D0 ttl=3D64 time=3D0.015 ms >> 64 bytes from 127.0.0.1: icmp_seq=3D1 ttl=3D64 time=3D0.049 ms >> 64 bytes from 127.0.0.1: icmp_seq=3D2 ttl=3D64 time=3D0.012 ms > > Fairly normal slowness for -current. > >> 64 bytes from [actual IP]: icmp_seq=3D0 ttl=3D64 time=3D0.019 ms >> 64 bytes from [actual IP]: icmp_seq=3D1 ttl=3D64 time=3D0.015 ms > > Are these with external hardware NICs? =A0Then 15 uS is excellent. =A0Bet= ter > than I've ever seen. =A0Very good hardware might be able to do this, but > I suspect it is for the local machine. =A0BTW, I don't like the times > been reported in ms and sub-uS times not being supported. =A0I sometimes > run Linux or cygwin ping and don't like it not supporting sub-mS times, > so that it always reports 0 for my average latency of 100 uS. > >>> After various tuning and bug fixing (now partly committed by others) I >>> get >>> improvements like the following on low-end systems with ffs (I don't us= e >>> zfs): >>> - very low end with 100Mbps ethernet: little change; bulk transfers >>> always >>> =A0went at near wire speed (about 10 MB/S) >>> - low end with 1Gbps/S: bulk transfers up from 20MB/S to 45MB/S (local >>> ffs >>> =A050MB/S). =A0buildworld over nfs of 5.2 world down from 1200 seconds = to 800 >>> =A0seconds (this one is very latency-sensitive. =A0Takes about 750 seco= nds on >>> =A0local ffs). >> >> Is this on 9.0-CURRENT, or RELENG_8, or something else? > > Mostly with 7-CURRENT or 8-CURRENT a couple of years ago. =A0Sometimes wi= th > a ~5.2-SERVER. =A0nfs didn't vary much with the server, except there were > surprising differences due to latency that I never tracked down. > > I forgot to mention another thing you can try easily: > > - negative name caching. =A0Improves latency. =A0I used this to reduce ma= keworld > =A0times significantly, and it is now standard in -current but not > =A0enabled by default. Have you also tried tuning via sysctl (vfs.nfs* ?) Thanks, -Garrett
Responsible Changed From-To: freebsd-bugs->freebsd-fs reclassify.
For bugs matching the following criteria: Status: In Progress Changed: (is less than) 2014-06-01 Reset to default assignee and clear in-progress tags. Mail being skipped
^Triage: I'm sorry that this PR did not get addressed in a timely fashion. By now, the version that it was created against is long out of support. Please re-open if it is still a problem on a supported version.